Effect of fasting on participation in clinical research among older people

Purpose: While many cohort studies have requested that participants fast prior to blood sampling, this may increase respondent burden, reduce response rates and add to the cost and complexity of the study. The objective of this study was to determine if fasting influences response rate to a health assessment and whether it impacts on time since last meal and lipid levels. Methods: A population-based sample of community dwelling people aged 50 and over was selected to undergo an in-home computer assisted personal interview (CAPI) followed by a physical assessment including blood sampling. Participants were randomly assigned to fast or not to fast prior to the physical assessment. Participants were asked to report time since last meal. A vacutainer system was used to obtain 18 mL of venous blood and lipid levels were measured. Results: Allocation to fasting did not affect participation with similar proportions of respondents (87/101 [86%] fasting versus 89/102 [87%] non-fasting; P = 0.8) attending the health assessments and (81/101 [80.2%] fasting versus 83/102 [81%] non-fasting; P = 0.9) having blood drawn. The average time since last meal for those randomized to fast was 11.4 hours and for those not asked to fast was 4.7 hours (P < 0.01). Lipid levels other than triglycerides did not vary significantly by fasting status. Fasting respondents were more likely to schedule a morning assessment (74% versus 34%, P < 0.01). Conclusions: Obtaining fasting samples from people aged 50 and over is feasible and does not negatively influence response rates but has implications for the scheduling of assessments.